SECT. 2] l.AHCK-SCALE INTERAt'TIONS 157 



o()inj)aris()n show n in Fig. 29b. Tlie transfer fornuilas relate the sea-air fluxes 

 to local, routinely measured meteorological jmrameters. Regardless of u!i- 

 certainties in their pro])orti()nality constants, the excellent agreement in the 

 months of maximum, mininuim and trends can hardly be accidental. It gives, 

 in fact, considerably increased confidence in the ability of both methods to 

 provide a useful quantitative picture of air-sea interaction and energetics. In 

 Fig. 29b, the magnitudes of total exchange Qs-\-Qe computed from the two 

 methods agree within 15% except in autumn when the energy-budget method 

 gives 30% higher transfer in October. Table XII compares the present results 

 for mean annual evaporation with those of other authors. 



Table XII 

 Comparison of Annual Evaporation Computations for the Caribbean Region 



Author Method Evaporation, 



cm/year 



Colon (1960) Energy budget 161 



Colon (1960) Formulas— Marine Atlas (U.S. Navy) 146 



Budyko (1956) Formulas— Atlas of the Heat Balance (U.S.S.R.) 138 



Jacobs (1951a) Formulas — U.S. Weather Bureau Climatic Charts 124 



Wiist (1936) Budget (Lat. 10°-20°N Atlantic) 146 



Separate transfer formula (20 and 21) evaluations of Qs and Qe were made 

 from synoptic ship data for the two special months of the meteorological study, 

 namely December, 1956, and January, 1957. The Bowen ratio came out exactly 

 10%. The sum Qe-\-Qs was somewhat higher than the climatic means computed, 

 as shown by the x's in Fig. 29b. Comparison of Figs. 29a and 18 is also very 

 good, showing the same magnitudes and seasonal march of the balance com- 

 ponents in climatically similar oceanic regions of opposite hemispheres. How- 

 ever, it should be pointed out that the methods of Budyko in obtaining Fig. 18 

 and of Colon in computing Table XI (from which Fig. 29a was made) are based 

 on the same type of radiation calculations and transfer formulas, the latter 

 with nearly identical coefficients. Later Australian work^ has suggested 

 somewhat larger coefficients in the transfer formulas than those of (20) and 

 (21), while, as we saw, the storage calculations of Pattullo (1957) suggest that 

 S could be larger (but not smaller) than found here. Clearly a larger 8 and Qe 

 are mutually incompatible in summer unless the radiation balance in that 

 season has been significantly (about 50%) underestimated. Therefore, a final 

 critique of all these joint budget studies rests upon and awaits improvement in 

 atmospheric radiation determinations, which are at present a real "bottleneck" 

 in much of meteorology and oceanography. 



1 Priestley (1959) has suggested an 88% increase in the coeflficient in equation (20) for 

 sensible heat flux, while an unpublished work by Swinbank argues for a 30% increase in 

 the coefficient in (21) for latent heat flux. 



